Should I Stay or Should I Go—Leveraging an Overlooked Feature of Click‐to‐Release Chemistry for Affinity Labeling

ABSTRACT

Click‐to‐release (C2R) chemistry is widely used for controlled payload liberation, yet the reactivity of its transient intermediates has remained largely unexplored. Here we show that an intermediate formed during the reaction of trans‐cyclooctene (TCO)‐appended protein‐targeting ligands with release tetrazines (rTzs), previously regarded as a byproduct rapidly quenched by water, can be exploited as an electrophile for ligand‐directed covalent protein labeling. The bioorthogonally generated dihydropyridazine methide (DHP‐methide) enables affinity‐directed protein labeling, resulting primarily in capture at surface‐exposed histidine residues proximal to the site of electrophile generation and revealing a histidine‐biased covalent modification pathway. Incorporation of functional handles into the rTz platform enables modular reporter installation via an orthogonal conjugation step. A fluorogenic leaving‐group design further allows direct optical reporting of the electrophile‐generating elimination event and real‐time monitoring of DHP‐methide formation. Selective labeling is retained in cell lysate and can be translated to a cellular context, as demonstrated by confocal imaging of surface‐expressed targets. These findings establish a bioorthogonally gated strategy for affinity labeling and reveal a previously unrecognized functional dimension of rTz–TCO click‐to‐release chemistry.

Click‐to‐release (C2R) chemistry is redirected from payload liberation toward ligand‐directed covalent protein labeling through a transient dihydropyridazine methide (DHP‐methide). This concept opens the way to dual‐functional systems combining covalent target engagement with controlled effector release.

1. Introduction

Covalent protein labeling is a cornerstone of chemical biology, proteomics, and drug discovery, enabling target identification, binding‐site mapping, and functional interrogation of biomolecular interactions in complex biological environments [1, 2, 3, 4, 5]. Among the growing repertoire of such methods, affinity labeling has emerged as a powerful strategy, in which a noncovalent ligand–protein interaction is converted into a stable covalent modification through proximity‐enabled nucleophile capture [6, 7]. In traditional affinity labeling schemes, ligands are equipped with constitutively reactive electrophilic warheads that undergo covalent capture upon target engagement. Pioneering examples by Hamachi and co‐workers demonstrated ligand‐directed systems employing electrophilic moieties such as tosylates [8], acyl imidazoles [9], and activated phenyl esters [10], which are positioned at the protein surface upon ligand binding and undergo covalent protein labeling via S_N2‐type substitution or addition–elimination mechanisms. These strategies leverage proximity to enhance selectivity while maintaining reduced intrinsic electrophilicity. Nevertheless, even moderately reactive electrophilic warheads may suffer from off‐target modification and lack of temporal control. Moreover, recent proteome‐wide analyses of electrophile reactivity have highlighted the complex and context‐dependent residue selectivity of electrophilic probes, underscoring the importance of strategies that provide spatial and temporal control over electrophile generation [11, 12]. To overcome these limitations, masked electrophile strategies that enable on‐demand activation have emerged. Among these, photoaffinity labeling (PAL) has become one of the most widely used approaches [13, 14, 15]. PAL employs light‐triggered activation of, that is, diazirines, benzophenones, or aryl azides to generate highly reactive species that covalently capture nearby nucleophiles. While powerful, PAL methods often suffer from low labeling efficiency, phototoxic side reactions and limited applicability in optically inaccessible environments. Although recent advances using visible light photocatalysis have mitigated some of these limitations [16, 17, 18], photochemical activation and radical pathways remain inherent features of these methods, potentially restricting their applicability in settings where light delivery is impractical.

Complementary to photochemical approaches, strategies for in situ electrophile generation by purely chemical means have also emerged. In particular, bioorthogonal reactions can be harnessed to generate reactive electrophilic species under physiological conditions. Recently, Dickinson et al. reported a chemically triggered proximity‐labeling strategy in which a subcellularly targeted esterase from Bacillus subtilis (BS2) activates masked enol ester probes to generate highly reactive bioorthogonal acylating agents for selective RNA labeling in living cells [19]. Prescher and coworkers demonstrated that phosphine‐mediated activation of cyclopropenones generates electrophilic ketene‐ylide intermediates that can be captured upon proximity‐induced assembly [20]. Diboron‐activated halogenated enamine N‐oxides release haloiminium species capable of ligand‐directed protein modification [21]. In a related work, tetrazine—trans‐cyclooctene (TCO) chemistry has been used to conditionally lock reversible cysteine conjugation rather than generate a new electrophile [22, 23]. Click‐to‐release reactions produce isocyanate electrophiles upon bioorthogonal reaction with cyclooctynes, representing an early example of electrophile generation through a click‐to‐release process and enabling biomolecular capture in proximity‐driven labeling settings [24]. Very recently, bioorthogonal retro‐Cope elimination reactions have been used to activate masked acrylamide and vinylsulfonamide warheads in living cells in the context of covalent inhibitors, enabling conditional covalent protein modification [25]. Despite these advances, the use of bioorthogonal reactions to generate transient small‐molecule electrophiles and harness them for ligand‐directed affinity labeling remains largely unexplored. Among bioorthogonal transformations, the inverse‐electron‐demand Diels–Alder (IEDDA) reaction between tetrazines and TCOs stands out as one of the fastest reactions known and has been widely applied in click‐to‐release (C2R) schemes for controlled payload liberation [26, 27, 28]. Mechanistic studies indicate that release tetrazine (rTz)–TCO bond cleavage generates a transient dihydropyridazine methide (DHP‐methide) intermediate that is typically regarded as rapidly quenched by water [26]. Following an unexpected covalent labeling observation, we sought to rationalize its origin and found that this intermediate does not need to be lost to water but can be retained and functionally exploited for protein labeling. The DHP‐methide formed during the rTz–TCO reaction can therefore serve as an electrophile for ligand‐directed covalent protein labeling when its formation is spatially confined by high‐affinity ligand binding and chemical precision. Covalent modification by the C2R‐generated electrophile is governed by the interplay of intrinsic nucleophile reactivity and the protein‐surface environment.

2. Results and Discussion

Our study originated from attempts to develop a bioorthogonally controlled fluorogenic proximity‐labeling strategy based on C2R chemistry [29]. We combined a release tetrazine rTz–caged coumarin‐derived quinone methide (QM) precursor (Cou‐rTz) with a TCO‐appended carbonic anhydrase (CA) sulfonamide ligand (SA‐TCO). The isopropyl‐substituted rTz scaffold was selected based on the mechanistic studies of Robillard and coworkers [26], whereas the equatorial 5‐hydroxy‐trans‐cyclooctene derivative was chosen following the work of Wu and coworkers demonstrating its favorable release properties in click‐to‐release chemistry [30]. Upon IEDDA activation and subsequent elimination, the system was designed to liberate a reactive QM pro‐fluorophore that was expected to be captured by nearby nucleophilic residues, yielding a covalently bound fluorescent coumarin (Figure 1). While this approach yielded the anticipated fluorogenic response, it failed to produce covalent labeling of the protein of interest via QM chemistry (Figure 1). Notably, however, mass spectrometric analysis consistently revealed the appearance of a new protein species indicative of covalent modification. Analysis of the mass shifts revealed a covalent adduct that could not be rationalized by QM capture but rather corresponded to modification by a pyridazine‐derived species (Figure 1). This unexpected observation prompted us to reexamine the click‐to‐release cascade. Based on the detailed mechanistic studies by Robillard et al. [26], we identified a dihydropyridazine‐ (DHP) methide species that is generated transiently upon an elimination step subsequent of the rTz–TCO cycloaddition, as the most plausible electrophilic species responsible for the observed covalent modification (Figure 2). Although this intermediate has previously been regarded as a short‐lived side product rapidly quenched by water, its apparent ability to modify the protein in an affinity‐dependent manner suggested that spatial confinement through ligand binding could enable productive nucleophilic capture.

FIGURE 1.

The expected and observed modification on bCA upon reaction of SA‐TCO and Cou‐rTz. Figure was created with BioRender.com.

FIGURE 2.

The general scheme for release from rTz upon TCO‐trigger and the structure of nucleophile‐captured and water‐quenched adducts. Figure was created with BioRender.com.

To gain deeper insight into the unexpectedly observed DHP‐methide‐mediated tagging, we examined the reaction from several perspectives. First, we evaluated the influence of the leaving group on the rTz unit by synthesizing carbamate (C‐rTz, Figure 2) and para‐nitrophenolate (pNP‐rTz) modified release tetrazines (Scheme S2).

An rTz containing fluoride as the leaving group was also considered, however, pronounced hydrolytic instability rendered this derivative unsuitable for labeling experiments. Comparative labeling experiments on bovine carbonic anhydrase II (bCA) using the respective rTzs and the ligand SA‐TCO revealed markedly different labeling kinetics despite comparable overall labeling efficiencies, affording 35% and 44% labeling yields for C‐rTz and pNP‐rTz, respectively (Figures S4, S5, and S6). Whereas pNP‐rTz afforded near‐maximal protein modification already after 2 h, labeling with C‐rTz proceeded substantially more slowly and reached comparable conversion only after prolonged incubation (Figures S4 and S5).

To determine whether the observed labeling yields are governed by the efficiency of the release step, we performed separate release experiments. To this end, a fluorogenic rhodol‐releasing analogue (Rh‐rTz, Scheme S4) was synthesized, in which the carbamate leaving group is connected through a self‐immolative linker to a quenched rhodol probe. Importantly, this design introduces a unique fluorogenic reporting feature intrinsically linked to the release process, as the same elimination cascade responsible for DHP‐methide generation simultaneously triggers fluorophore liberation. Upon tetrazine–TCO reaction, release from Rh‐rTz was followed by fluorescence increase and LC–MS analysis (Figure S2), while release from pNP‐rTz was monitored by UV–vis spectroscopy at the absorption wavelength of the liberated nitrophenolate (Figure S2). Both rTz variants reached similar overall release yields of approximately 73%, consistent with literature reports [26], however, nitrophenolate release was complete within ∼90 min, whereas C‐rTz required ∼10 h. Despite the pronounced kinetic differences, protein labeling efficiencies were comparable, indicating that modification efficiency is governed not simply by release conversion or rate, but rather by the competition between nucleophilic trapping of the DHP‐methide and its quenching by water. Although pNP‐rTz displayed more favorable release kinetics, this scaffold was less stable and failed to provide derivatives suitable for further functionalization. Considering the aqueous stability of C‐rTz (Figure S1) together with its synthetic accessibility and modularity, subsequent studies focused on this scaffold despite its slower release kinetics.

We next attempted to quantify the kinetics of nucleophilic capture of the generated DHP‐methide. However, consistent with the mechanistic studies of Robillard and coworkers [26], direct kinetic analysis proved challenging because DHP‐methide generation during the release step constitutes the rate‐determining process, while the formed electrophile reacts rapidly with either proximal nucleophiles or water. Nevertheless, the unique fluorogenic reporting capability of the Rh‐rTz scaffold provided indirect mechanistic insight into the temporal relationship between DHP‐methide generation and protein labeling. In parallel with the fluorescence increase resulting from rhodol liberation, time‐dependent protein modification yields were determined under identical conditions (Figure 3). Although fluorophore liberation additionally requires self‐immolation of the linker, related self‐immolative systems undergo fragmentation within minutes [40], whereas release from Rh‐rTz proceeds over several hours. Consistent with this difference in timescale, LC–MS monitoring did not reveal accumulation of the primary release product preceding self‐immolation. These observations indicate that self‐immolation is unlikely to be rate‐limiting and that the fluorescence increase predominantly reflects the temporal progression of the slower C2R release process associated with DHP‐methide generation. Although qualitative in nature, comparison of the fluorescence and protein‐modification profiles therefore provides information on the temporal relationship between electrophile generation and subsequent protein labeling.

FIGURE 3.

Fluorogenic reporting of protein labeling: labeling efficiency on bCA with SA‐TCO and C‐rTz, measured with intact MS (c(SA‐TCO) = 50 µM, c(C‐rTz) = 100 µM, c(bCA) = 1 mg/ml in PBS containing 2% DMSO, pH = 7.4 (left panel)), and fluorescence increase of released rhodol at 565 nm from Rh‐rTz upon reaction with SA‐TCO in the presence of bCA (reaction: c(SA‐TCO) = 50 µM, c(Rh‐rTz) = 100 µM, c(bCA) = 1 mg/ml in PBS containing 2% DMSO, pH = 7.4; fluorescence measurement: c(SA‐TCO) = 0.5 µM, c(Rh‐rTz) = 1 µM, c(CA) = 10 µg/ml in PBS, pH = 7.4; right panel).

Next, we evaluated the specificity and selectivity of the C2R mediated affinity labeling method. To this end, a TCO appended non‐binding bCA ligand S10 was synthesized and added to bCA. Following addition of C‐rTz, no modification was observed even after 24 h of reaction time (Figure S8). Likewise, no modification was observed when we applied the TCO‐modified CA‐ligand (SA‐TCO) to trypsin inhibitor (TI) or β‐lactoglobulin A (Figure S10). These observations supported the ligand‐directed affinity targeting of the method. We additionally performed a post‐labeling displacement with acetazolamide, a ligand that exhibits stronger binding affinity than SA‐TCO. Following bioorthogonally triggered modification of bCA with SA‐TCO and C‐rTz, acetazolamide was added in excess (Figure S11). We observed no change in the mass spectrum, further confirming covalent modification. To identify the labeling site and validate the covalent nature of the modification, the labeled bCA was subjected to trypsin digestion, and the resulting peptide fragments were analyzed by LC–MS/MS. The presence of a characteristic reporter ion enabled identification of the modified peptide. We concluded that in bCA the DHP‐methide modifies a histidine residue near the N‐terminus (Figure S12).

To further assess the generality of the approach beyond carbonic anhydrase, we extended the C2R‐based affinity‐labeling strategy to bromodomain‐containing proteins. Recombinantly expressed second bromodomain of human BRD2 (BD2) and first bromodomain of human BRD4 (BD1) were incubated with a TCO‐appended (+)‐JQ1 ligand (JQ1‐TCO, Table 1), followed by treatment with C‐rTz. Efficient covalent labeling was observed within 10 h for both BD2 and BD1 at the intact‐protein level, affording labeling efficiencies of 62% and 46%, respectively (Figure S7). Despite repeated attempts, residue‐level site assignment by LC–MS/MS could not be obtained for BD2. In the case of BD1, modification was observed either at lysine (K70) or tyrosine (Y18) residues, consistent with the absence of suitably positioned histidine residues in proximity to the binding site (Figure S14). It should be noted that both bromodomains required slightly basic conditions (pH 8.5) for solubilization, which may contribute to the observed residue reactivity pattern. In contrast, TCO‐appended non‐binding controls, S13 based on (‐)‐JQ1 or S10 failed to produce covalent modification of either BD2 or BD1, confirming strict ligand dependence (Figure S8).

TABLE 1.

The structure of TCO‐modified ligands and the modification yields on corresponding proteins. ^a

Structure of TCO‐ligands	Target protein	Labeling efficiency
	bCA	35% b
	BD2	62% c
	BD1	46% c
	eDHFR	18% b

^a Labeling efficiencies were measured after 10 h. c(TCO) = 50 µM, c(C‐rTz) = 100 µM.

^b In PBS (pH = 7.4, 2% DMSO).

^c In sodium borate buffer (pH = 8.5, 2% DMSO).

We also applied the C2R‐generated DHP‐methide platform to Escherichia coli dihydrofolate reductase (eDHFR) using a corresponding TCO‐appended ligand (TRIM‐TCO, Table 1), where covalent modification at a histidine residue was observed at pH 7.4 (His67, Figure S13). These findings indicate that DHP‐methide–mediated affinity labeling is ligand‐directed and compatible with multiple protein classes and ligand scaffolds. In systems amenable to residue‐level mapping, histidine emerged as the predominant site of modification when suitably positioned and under physiological conditions, as observed for bCA and eDHFR. Because histidine‐selective covalent modification remains relatively underexplored compared to cysteine‐ or lysine‐targeting electrophiles [13, 31], the repeated identification of histidine as the modification site prompted us to directly examine the intrinsic residue selectivity of the C2R‐generated electrophile. To this end, quantitative analysis of reactions with surrogate N‐acetylated amino acids (Tyr, Ser, Lys, Cys, and His) at pH 7.4 were performed, which revealed marked differences in modification levels (Figure 4). Under these conditions, histidine exhibited nearly two orders of magnitude higher apparent reactivity than cysteine, the next most reactive residue, whereas the remaining nucleophiles showed only minimal conversion. At the pH of BD modification studies (pH 8.5) the overall reactivities increased, most notably for cysteine, consistent with enhanced thiolate formation (Figure S3). Yet, histidine remained the most reactive residue under both conditions. This pronounced preference likely arises from a combination of residue protonation state, intrinsic nucleophilicity, adduct stability, and the local protein environment. At pH 7.4, a substantial fraction of histidine side chains is present as the neutral, nucleophilic imidazole, whereas cysteine remains predominantly protonated despite the high intrinsic nucleophilicity of its thiolate form. At the protein level, these intrinsic reactivity differences are further shaped by residue accessibility, the local microenvironment, and ligand‐mediated spatial confinement. Importantly, surface‐exposed cysteine residues are comparatively rare and often engaged in disulfide bonds, whereas histidine residues are more frequently accessible. Together, these factors account for the predominance of histidine modification observed in our protein‐labeling experiments under physiological conditions. Importantly, however, these results indicate that DHP‐methide‐mediated labeling is histidine‐biased rather than histidine‐specific. Histidine emerged as the preferred site of covalent capture when suitably positioned in proximity to the ligand‐binding site, as observed for bCA and eDHFR under physiological conditions. The absence of a suitably positioned histidine does not preclude labeling, as demonstrated for BD1, where lysine and tyrosine residues were modified. Thus, the applicability and efficiency of the approach are expected to depend not on the presence of histidine alone, but more broadly on the spatial distribution, accessibility, and local effective concentration of nucleophilic residues surrounding the ligand‐binding site. Because productive nucleophilic capture competes with rapid hydrolytic quenching of the DHP‐methide, the local protein environment is expected to be a key determinant of labeling efficiency.

FIGURE 4.

Amino acid selectivity of the in situ generated DHP‐methide, confirmed with intrinsic reactivity studies of C‐rTz and SA‐TCO in the presence of excess N‐acetylated surrogate amino acids and ion chromatogram of the reaction products carried out with N‐acetyl‐histidine. HPLC‐MS analysis was conducted in the presence of 200 µM indoprofen as an internal standard, c(C‐rTz) = 0.5 mM, c(SA‐TCO) = 2.5 mM, c(N‐acetylated amino acid) = 5 mM in PBS (containing 10% organic solvent).

While these experiments rationalize the observed residue selectivity, they do not address the fate of the fraction of DHP‐methide that does not undergo productive covalent capture. To gain insight into the fate of the generated DHP‐methide, the reaction was carried out in the presence of N‐acetyl‐histidine and the resulting product distribution was analyzed by LC–MS (Figure 4). In addition to the expected histidine adduct, the corresponding water‐quenched DHP‐OH species was identified, together with oxidation products derived from DHP‐OH, residual unreleased click adducts, and dead‐end products. These observations indicate that DHP‐methide capture by nucleophiles competes with several parallel pathways, with hydrolytic quenching appearing to predominate under these homogeneous aqueous conditions. Direct determination of the intrinsic lifetime of the DHP‐methide is complicated by its comparatively slow, rate‐limiting generation during the release step, which prevents independent kinetic resolution of its subsequent consumption. Accordingly, the present data do not permit assignment of a quantitative lifetime. Importantly, in the affinity‐labeling system described here, the DHP‐methide remains covalently tethered to the affinity ligand and therefore does not diffuse away as an independently released reactive species. Its spatial reach is instead constrained by the geometry of the ligand–protein complex and linker flexibility, while productive capture depends on the accessibility and effective local concentration of proximal nucleophilic residues. Thus, rather than being characterized by a diffusion‐dependent labeling radius, the spatial resolution of this system is expected to be determined primarily by the local architecture of the ligand‐bound protein and by competition between proximal nucleophilic capture and hydrolytic quenching.

To demonstrate the modularity of the rTz platform, we synthesized azide‐functionalized derivative, N₃‐rTz (Scheme S3) to enable post‐labeling conjugation. Following C2R‐mediated modification, the azide‐bearing protein conjugates were treated with DBCO‐linked fluorescent probes (AF488, Cy3 or Cy5, Figure S15), resulting in efficient formation of the corresponding fluorescently labeled proteins as seen on SDS‐PAGE gels (Figures S16–S19). To further illustrate the utility of the two‐step C2R‐based labeling strategy, we next investigated selective post‐tagging in defined protein mixtures. Experiments were performed using mixtures containing bCA, BD2 and trypsin inhibitor (TI). In two parallel experiments, a single TCO‐modified ligand specific to target one protein at a time was introduced, followed by treatment with N₃‐rTz and subsequent conjugation with a DBCO‐functionalized fluorophore. Gel analysis revealed selective fluorescent labeling of the intended target protein in each case, demonstrating that ligand‐controlled electrophile generation enables precise addressing of individual proteins within a multi‐component system (Figure 5). These results indicate that C2R‐mediated electrophile generation remains effective and ligand‐directed in a defined multicomponent protein mixture.

FIGURE 5.

SDS‐PAGE analysis of selective protein labeling in defined protein mixtures (left) and bCA labeling in spiked HEK293T cell lysate (right). Labeling of bCA and BD2 with ligands SA‐TCO or JQ1‐TCO, respectively, and tetrazine N₃‐rTz was carried out at 25°C for 24 h, followed by fluorescent labeling with added DBCO dyes for additional 24 h. Labeling in protein mixture: c(TCO) = 50 µM, c(N₃‐rTz) = 100 µM, c(DBCO dyes) = 200 µM, c(bCA, BD2 and TI) = 1 mg/ml in sodium borate buffer containing 2% DMSO (pH = 8.5), 15% resolving gel was used. Labeling in cell lysate: c(SA‐TCO) = 3 µM, c(N₃‐rTz) = 3 µM, c(AF488‐DBCO) = 3 µM, c(bCA) = 33 µg/ml (1 µM) in buffer containing 2% DMSO, 10% resolving gel was used.

To evaluate selectivity in a more complex proteomic environment, we next performed labeling experiments in HEK293T cell lysate spiked with bCA. Following addition of SA‐TCO and N₃‐rTz and subsequent conjugation with AF488‐DBCO, SDS‐PAGE analysis revealed selective fluorescent labeling of bCA at a final target concentration of 1 µM, with minimal detectable off‐target labeling from endogenous lysate proteins (Figure 5). Selective target labeling remained detectable when the concentration of spiked bCA was reduced to 0.3 µM (Figure S20). These results demonstrate that ligand‐directed DHP‐methide capture can retain substantial target selectivity even in a complex cellular protein background.

Next, we examined whether the C2R‐based labeling and post‐tagging strategy can be translated to a cellular setting. Cytotoxicity studies confirmed that none of the applied reagents induced detectable toxicity (Figure S22). Live HEK293T cells expressing BD2 fused to interleukin‐2 receptor alpha and fluorescent reporter protein EGFP (BD2‐IL2Rα‐EGFP) on their surface were treated with JQ1‐TCO and N₃‐rTz, followed by labeling with Cy5‐DBCO. Confocal microscopy images revealed clear and ligand‐dependent fluorescence that overlapped with the EGFP reporter signal, confirming that C2R‐mediated electrophile generation and subsequent post‐tagging can be successfully implemented for protein labeling in live‐cell settings (Figure 6A–C). At the same time, minimal fluorescence was observed in control experiments lacking the TCO‐ligand and the rTz component (Figure 6G–I), upon replacement of JQ1‐TCO with a non‐binding ligand (S10) (Figure 6D–F), or when non‐transfected cells were treated with JQ1‐TCO and N₃‐rTz (Figure 6J–L), confirming the ligand‐dependent nature of the labeling process.

FIGURE 6.

Confocal microscopy images of HEK293T cells. (A–C): Cells expressing BD2‐IL2Rα‐EGFP on their surface were treated with JQ1‐TCO (3 µM), N₃‐rTz (4.5 µM) and Cy5‐DBCO (4.5 µM); (D–F): cells expressing BD2‐IL2Rα‐EGFP were treated with non‐binding S10 (3 µM), N₃‐rTz (4.5 µM) and Cy5‐DBCO (4.5 µM); (G–I): cells expressing BD2‐IL2Rα‐EGFP were treated with Cy5‐DBCO (4.5 µM); (J–L): cells not expressing BD2‐IL2Rα‐EGFP were treated with JQ1‐TCO (3 µM), N₃‐rTz (4.5 µM) and Cy5‐DBCO (4.5 µM). Scale bar: 20 µm.

3. Conclusion

Collectively, these findings reveal an overlooked functional dimension of release tetrazine–trans‐cyclooctene click‐to‐release chemistry, in which a transient electrophilic intermediate can be harnessed for affinity‐directed covalent protein labeling with a characteristic histidine‐biased capture profile. Rather than being obligatorily quenched by water, the dihydropyridazine methide can be intercepted when its formation is spatially confined by high‐affinity ligand binding, enabling selective and irreversible protein modification under bioorthogonal control. Binding specificity was validated using non‐binding ligands and off‐target proteins, while the covalent nature of modification was confirmed by displacement experiments and MS/MS identification of modified residues at the protein surface. Mechanistically, covalent capture by the C2R‐generated electrophile arises from the interplay between intrinsic residue reactivity and protein‐surface context, resulting primarily in a histidine‐biased modification pattern under affinity‐directed conditions.

A distinctive feature of this platform is that electrophile generation itself can be directly reported when the rTz leaving group is designed as a fluorogenic reporter. In such systems, the same elimination cascade that produces the DHP‐methide also triggers self‐immolative fluorophore release, enabling direct optical monitoring of electrophile formation independently of downstream protein conjugation. This built‐in reporting capability links signal generation to the underlying C2R mechanism and provides real‐time insight into the electrophile‐generating step during protein labeling.

Beyond mechanistic insight, the modularity of the rTz platform enables downstream functionalization of labeled proteins via orthogonal chemistry. Selective labeling was retained in defined protein mixtures and cell lysate and could be extended to a cellular context as demonstrated by confocal imaging of surface‐expressed targets following post‐labeling fluorophore installation. Given the tunability of rTz scaffolds and ongoing advances aimed at accelerating the release step in tetrazine–TCO click‐to‐release chemistry, this strategy has the potential to evolve into a versatile framework for chemically gated affinity‐labeling systems. Its purely chemical activation mode may also prove advantageous in settings where optical control is impractical and complements existing photochemical and bioorthogonal activation strategies.

More broadly, the present work extends the conceptual scope of click‐to‐release chemistry from a technology for payload liberation to one capable of ligand‐directed installation of persistent molecular function at target proteins. Beyond serving as a complementary affinity‐labeling strategy, this previously unrecognized capability opens opportunities for dual‐function C2R architectures in which a single bioorthogonal activation event converts a reversible ligand into a covalent binder while simultaneously liberating a biologically active payload. Such systems could enable synergistic modes of action that combine durable target engagement with controlled effector release. The feasibility of such systems is currently being explored and the results will be reported in due course.

Author Contributions

Dóra Kern: conceptualization, methodology, validation, investigation, supervision, writing – original draft, writing – review and editing. Krisztina Németh: conceptualization, investigation, methodology, supervision. Tímea Imre: investigation, validation. József Simon: investigation, validation. Kitti Meggyes: investigation. Gergő Bárdos: investigation. László Petri: investigation, validation. Márton Bojtár: conceptualization, methodology. Attila Kormos: conceptualization, methodology, validation, investigation, writing – review and editing, supervision. Péter Kele: conceptualization, investigation, funding acquisition, writing – review and editing, methodology, validation, supervision.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

The authors have cited additional references within the Supporting Information [32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46].

Supporting File 1: anie74499‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are openly available in chemrxiv at https://doi.org/10.26434/chemrxiv‐2025‐gmghk/v3.